基于Grad-CAM与KL损失的SSD目标检测算法  

SSD Object Detection Algorithm Based on KL Loss and Grad-CAM

作　　者：侯庆山 邢进生[1] HOU Qing-shan;XING Jin-sheng(School of Mathematics and Computer Science,Shanxi Normal University,Linfen,Shanxi 041004,China)

机构地区：[1]山西师范大学数学与计算机科学学院,山西临汾041004

出　　处：《电子学报》2020年第12期2409-2416,共8页Acta Electronica Sinica

基　　金：山西省软科学基金资助项目(No.2011041033-03)。

摘　　要：鉴于Single Shot Multibox Detector(SSD)算法对中小目标检测时会出现漏检甚至错检的情况,提出一种改进的SSD目标检测算法,以提高中小目标检测的准确性.运用Gradient-weighted Class Activation Mapping(Grad-CAM)技术对检测过程中的细节作可视化处理,并以类激活图的形式呈现各检测层细节,分析各检测层的类激活图发现SSD算法中待检测目标的错检以及中小目标的漏检现象与回归损失函数相关.据此,采用Kullback-Leibler(KL)边框回归损失策略,利用Non Maximum Suppression(NMS)算法输出最终预测框.实验结果表明,改进算法相较于已有检测算法具有更高的准确率以及稳定性.Considering that the single shot multibox detector(SSD)algorithm will be missed or even false when it is used to detect the small and medium-sized objects,an improved SSD object detection algorithm is proposed to improve the accuracy of small and medium-sized objects detection.The details in the detection process are visualized with gradient-weighted class activation mapping(Grad-CAM)technology,and the details of each detection layer are shown in the form of class activation maps.Then it is noted that the phenomenon of the false or missed detection of the objects to be detected on small and medium-sized objects in the SSD algorithm is related to the regression loss function.Accordingly,Kullback-Leibler(KL)border regression loss strategy is adopted and non maximum suppression(NMS)algorithm is used to output the final prediction boxes.Experimental results show that compared with the existing detection algorithms,the improved algorithm in this paper has higher accuracy and stability.

关 键 词：目标检测 可视化 类激活图 Grad-CAM SSD KL损失 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]